文献速递神经技术与安全CT

基于计算机断层扫描的深度学习驱动脑卒中多分类:迈向更高诊断精度的一步

Deep learning-driven multi-class classification of brain strokes using computed tomography: A step towards enhanced diagnostic precision.
PubMed 文献 · Eur J Radiol · 2025-04-11

这项回顾性双中心研究纳入250例患者、2017至2022年间采集的8186幅CT图像,基于Expanded ResNet101框架构建两步法深度学习模型进行脑卒中预测与分类。模型01内部测试准确率99.6%、精确率99.4%、F1分数99.6%,模型02分别为99.2%、98.8%、99.1%;外部验证中专家评估的准确率分别为78.6%(95%CI 0.73,0.83; P<0.001)和60.2%(95%CI 0.48,0.70; P<0.001)。作者认为在扩大队列并纳入更多放射学模仿病变后,这些模型可支持临床预后判断与决策。

要点

  • 两步法深度学习模型在内部测试集上准确率分别为99.6%和99.2%。
  • 专家评估的外部验证准确率模型01为78.6%、模型02为60.2%。
  • 作者认为需扩大队列并纳入更多放射学模仿病变后模型才可能支持临床决策。

核心结论

  • 深度学习模型在预测脑卒中患者结局方面表现出高准确率、精确率和F1分数。
  • 作者认为,在扩大队列并纳入更多放射学模仿病变后,这些模型可能支持临床医生的预后判断与决策。

研究设计与样本

  • 回顾性双中心研究,纳入250例患者,2017至2022年间采集8186幅CT图像。
  • 以Expanded ResNet101深度学习框架构建两步法模型:模型01训练用6386幅、测试与验证用900幅;模型02训练用1619幅、测试与验证用600幅。
  • 采用混淆矩阵评估模型性能,并用独立数据集进行外部验证,外部验证由一名专家和两名外部成员完成。
  • 计算总体准确率、置信区间、Cohen Kappa值及McNemar检验P值。

诊断标准与关键数值

  • 模型01平均准确率99.6%、精确率99.4%、F1分数99.6%;模型02平均准确率99.2%、精确率98.8%、F1分数99.1%。
  • 专家评估的外部验证准确率:模型01为78.6%(95%CI 0.73,0.83; P<0.001),模型02为60.2%(95%CI 0.48,0.70; P<0.001)。

临床应用与阅片要点

  • 原文未据此给出推荐。

局限性

  • 作者指出需扩大队列并纳入更多放射学模仿病变后,模型才可能支持临床决策。

以上解读依据论文摘要(全文未开放获取或许可不允许翻译转载)。

英文原文(摘要)· 点击展开

OBJECTIVE: To develop and validate deep learning models leveraging CT imaging for the prediction and classification of brain stroke conditions, with the potential to enhance accuracy and support clinical decision-making.

MATERIALS AND METHODS: This retrospective, bi-center study included data from 250 patients, with a dataset of 8186 CT images collected from 2017 to 2022. Two AI models were developed using the Expanded ResNet101 deep learning framework as a two-step model. Model performance was evaluated using confusion matrices, supplemented by external validation with an independent dataset. External validation was conducted by an expert and two external members. Overall accuracy, confidence intervals, Cohen's Kappa value, and McNemar's test P-values were calculated.

RESULTS: A total of 8186 CT images were incorporated, with 6386 images used for the training and 900 datasets for testing and validation in Model 01. Further, 1619 CT images were used for training and 600 datasets for testing and validation in Model 02. The average accuracy, precision, and F1 score for both models were assessed: Model 01 achieved 99.6 %, 99.4 %, and 99.6 % respectively, whereas Model 02 achieved 99.2 %, 98.8 %, and 99.1 %. The external validation accuracies were 78.6 % (95 % CI: 0.73,0.83; P < 0.001) and 60.2 % (95 % CI: 0.48,0.70; P < 0.001) for Models 01 and 02 respectively, as evaluated by the expert.

CONCLUSION: Deep learning models demonstrated high accuracy, precision, and F1 scores in predicting outcomes for brain stroke patients. With larger cohort and diverse radiologic mimics, these models could support clinicians in prognosis and decision-making.

来源:Eur J Radiol,查看原文。版权归原作者所有。

引用

  1. Kulathilake CD, Udupihille J, Abeysundara SP, et al. Deep learning-driven multi-class classification of brain strokes using computed tomography: A step towards enhanced diagnostic precision. Eur J Radiol. 2025. PMID: 40252282. doi:10.1016/j.ejrad.2025.112109

供执业医师工作参考,不替代临床判断;数值与分级以指南原文为准。